Class 12: ggplot2 Continued

Tidyverse: a process supported by R packages

Install the tidyverse
This week: visualization, transformation, exploration
Rectangular data: collections of values that are each associated with a variable and an observation.
Focus on data exploration and hypothesis generation.

Reference material: tidyverse, ggplot2, dplyr, slidy

Classes 11-19: The Tidyverse tidyverse Wickham’s Tidy data article

And, on Monday …

Exploratory Data Analysis

Monday (19 Feb 2018) EDA

ggplot2: let’s get started

We’re using the mpg dataset that comes with ggplot2.
Use data() to review available datasets and ?mpg to access the doc for the mpg dataset.

## Warning: replacing previous import by 'tidyr::%>%' when loading 'broom'

## Warning: replacing previous import by 'tidyr::gather' when loading 'broom'

## Warning: replacing previous import by 'tidyr::spread' when loading 'broom'

The mpg dataset reports on 234 cars on 11 variables:
manufacturer, model, displ, year, cyl, trans, drv, cty, hwy, fl, class

A small sample (note the table produced with knitr::kable())

kable(mpg[c(2,30,50,100,100),])

manufacturer	model	displ	year	cyl	trans	drv	cty	hwy	fl	class
audi	a4	1.8	1999	4	manual(m5)	f	21	29	p	compact
chevrolet	k1500 tahoe 4wd	5.3	2008	8	auto(l4)	4	11	14	e	suv
dodge	dakota pickup 4wd	3.7	2008	6	auto(l4)	4	14	18	r	pickup
honda	civic	1.6	1999	4	manual(m5)	f	28	33	r	subcompact
honda	civic	1.6	1999	4	manual(m5)	f	28	33	r	subcompact

ggplot() The basic idea

ggplot(data = ) + (mapping = aes())

ggplot(data = mpg) makes an empty graph.

Add layers to it with GEOM_FUNCTIONS.
List of GEOMs Also, see
Each geom takes mapping argument which uses aes() to map the x and y axes.
Add additional dimensions by mapping with color, shapes, size to other variables

First Plot

ggplot(data = mpg) + 
  geom_point(mapping = aes(x = displ, y = hwy))

Now. Make your first ggplot: hwy vs. cyl.

hwy vs. cyl

p <- ggplot(data = mpg) 

p + geom_point(mapping = aes(x = cyl, y = hwy))

Add a new dimension

The gas mileage data has more dimensions then just hwy, cyl, and displ.
Let’s add class as a dimension in our plot by mapping class to color.

p + geom_point(mapping = aes(x =  displ, y = hwy, color = class), size = 2)

ggplot has other mapping options: shape, size, alpha,

Go back to cyl

p + geom_point(mapping = aes(x =  displ, y = hwy, color = cyl), size = 2)

Problem: cyl is a continuous variable that takes only 4 values: 4, 5, 6, 8

Fix it

mpg$cyl <- as.factor(mpg$cyl)

p <- ggplot(data = mpg)
p +  geom_point(mapping = aes(x =  displ, y = hwy, color = cyl))

Setting color or shape or size, but not mapping?

p + geom_point(mapping = aes(x = displ, y = hwy), color = "red", shape = 22, size = 1)

Facets for visual Comparisons

p + geom_point(mapping = aes(x = displ, y = hwy)) +
  facet_wrap( ~ class, nrow = 2)

##Graphs that calculated new values to plot {.build}

smooth
Bar
Box

Geometric objects: smooth

p +  geom_point(mapping = aes(x = displ, y = hwy)) +
     geom_smooth(mapping = aes(x = displ, y = hwy))

Smooths for subsets

p +  geom_smooth(mapping = aes(x = displ, y = hwy, linetype = drv))

Bar charts: transformations

ggplot(data = diamonds) + geom_bar(mapping = aes(x = cut))

prop

ggplot(data = diamonds) +
  geom_bar(mapping = aes(x = cut, y = ..prop..,group=1))

Grouped bar chart data

source

# create a dataset
library(knitr)
library(kableExtra)
specie <- c(rep("sorgho" , 3) , rep("poacee" , 3) , rep("banana" , 3) , rep("triticum" , 3) )
condition <- rep(c("normal" , "stress" , "Nitrogen") , 4)
value <- abs(rnorm(12 , 0 , 15))
data <- data.frame(specie,condition,value)
options(digits = 2)
a <- kable(data, "html")
column_spec(a, 1:2, width = "1in")

specie	condition	value
sorgho	normal	13.6
sorgho	stress	13.9
sorgho	Nitrogen	58.3
poacee	normal	19.6
poacee	stress	1.3
poacee	Nitrogen	4.3
banana	normal	19.5
banana	stress	9.7
banana	Nitrogen	21.9
triticum	normal	9.9
triticum	stress	1.4
triticum	Nitrogen	14.2

Grouped data barchart “dodge”

# Grouped
ggplot(data, aes(fill=condition, y=value, x=specie)) + 
    geom_bar(position="dodge", stat="identity")

Grouped data barchart “stacked”

# Stacked
ggplot(data, aes(fill=condition, y=value, x=specie)) + 
    geom_bar( stat="identity")

Grouped data barchart “fill”

# Stacked Percent
ggplot(data, aes(fill=condition, y=value, x=specie)) + 
    geom_bar( stat="identity", position="fill")

Color Brewer

library(RColorBrewer)
display.brewer.all(n=10)

## Color Brewer 2

display.brewer.pal(10,"Set1")

## Warning in display.brewer.pal(10, "Set1"): n too large, allowed maximum for palette Set1 is 9
## Displaying the palette you asked for with that many colors

Using ColorBrewer in a bar chart

# color with RcolorBrewer
ggplot(data, aes(fill=condition, y=value, x=specie)) + 
    geom_bar( stat="identity", position="fill") +    
    scale_fill_brewer(palette = "Set1")

Facet instead of grouping

# Faceting
ggplot(data, aes(y=value, x=specie, color=specie, fill=specie)) + 
    geom_bar( stat="identity") +    
    facet_wrap(~condition) + theme(legend.position = "none")   # show w and w/o legend

Reorder the bars 1

order of bars

Name <- c( 'Juan','Michael','Andrea','Charles','Jonás','Juan','Donata','Flavia' )
City <- c('Madrid','New York','Madrid','Liverpool','Madrid','Buenos Aires','Rome','Liverpool')
City.Id <- c(1,2,1,3,1,4,5,3)
df = data.frame( Name,City,City.Id )

a <- ggplot( df,aes( x = City, text=paste("City.Id=",City.Id)) ) +geom_bar()

a

fct_ inorder, infreq

From:library(forcats)

vec <- c("a","b","c","d","a","c","d","b","d","c","d","c","a","b")

vec

##  [1] "a" "b" "c" "d" "a" "c" "d" "b" "d" "c" "d" "c" "a" "b"

fvec <- as.factor(vec)

fvec

##  [1] a b c d a c d b d c d c a b
## Levels: a b c d

fct_inorder(fvec)

##  [1] a b c d a c d b d c d c a b
## Levels: a b c d

fct_infreq(fvec)

##  [1] a b c d a c d b d c d c a b
## Levels: c d a b

Reorder the bars 2